Ship multi-target detection method based on rotation area extraction

Abstract

The invention discloses a ship multi-target detection method based on rotation area extraction, and the method comprises the steps of collecting a remote sensing image containing a ship target, and cutting the remote sensing image into a set size; carrying out rotation rectangle marking on ship the targets contained in the remote sensing images respectively, and taking the remote sensing images marked with rotation rectangles as a training set and a test set; inputting the training set into a ship multi-target detection model for training; inputting the remote sensing image to be detected intothe trained ship multi-target detection model for detection to obtain a preliminary detection result comprising a ship target pre-selection frame and the corresponding confidence; and calculating a rotation intersection-union ratio of the preselected frame with the highest confidence coefficient and other preselected frames, and deleting the preselected frame with the rotation intersection-unionratio greater than a set threshold to obtain a final detection result. By predicting the minimum area circumscribed rectangle of the ship target, the position, the length, the width and the axial information of the ship target are accurately outputted, the multi-target detection is realized, and the detection performance is remarkably improved.

Description

technical field [0001] The invention belongs to the technical field of remote sensing image processing, and more specifically relates to a ship multi-target detection method based on rotation area extraction. Background technique [0002] In recent years, satellite remote sensing technology has developed rapidly, and batches of remote sensing satellites with high resolution and low revisit cycle continue to emerge. With the gradual improvement of space-based observation systems, remote sensing technology has been widely used in all aspects of our daily life , has a major impact on civilian and military fields such as sea surface temperature analysis, fishery assistance, environmental and agricultural monitoring, and situational awareness on the sea surface battlefield. Remote sensing technology has also received more and more attention from various countries in the world, and even raised it to a strategic height , as the strategic goal of the country's long-term stable devel...

AI Technical Summary

Problems solved by technology

However, the existing rotating region extraction algorithms have a variety of definitions for rotating objects, all of which have disadvantages such as complex definitions, difficult model construction, and slow prediction speed.

Images